Recognizing the handwritten characters and converting them into machine-editable text is very tedious due to the diversity of writing styles and character patterns. Extracting data from images and identifying the characters becomes more complicated when a language consists of compound structures and characters, such as Bengali. There has been a lack of programs for recognizing Bengali scripted basic and com-plex numeric signs and letters with high accuracy. This paper develops a novel approach to extracting and identifying Bengali handwritten primary characters, digits, and primarily used compound characters. In this proposed model, an image containing Bengali handwritten text takes as input and processed. Then processed images are segmented into lines and characters. The features are extracted from segmented characters and recognized using a Convolutional Neural Network (CNN). The CNN obtains 98.23% accuracy in the training dataset and 96.02% in the validation dataset. Apart from that, the proposed model has gained 89.6% precision and 92.6% recall scores on scanned image data.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.